Ideally, u hv to ensure that you have a validated instrument with established construct validity before examining mean differences using t-tests or other inferential statistical tests. This helps ensure that any mean differences observed are based on a valid measurement of the latent factors or constructs of interest, and increases the confidence in the results of your hypothesis testing.

Hello Yumna,

in the course of any preliminary exploratory analysis, you can do that of course. I guess you have some groups in mind? If group differences become apparent, you could consider setting up the CFA model as a multiple-group model from the start.

Best,

Holger

Thankyou Holger Steinmetz ; really helpful. Yes the groups in mind differ by gender and marital status. I followed setting up CFA first .